OWPT Poster Session

[OWPTp-02] Improvement of communication quality by adaptive optics using Convolutional Neural Network

*Monami Teranishi¹, Kayo Ogawa¹ (1. Japan Women’s University)

In this study, we investigate the improvement of communication quality by adaptive optics using convolutional neural networks (CNN). The CNN predicts turbulent phase distribution from intensity distribution images after atmospheric turbulence. This conjugate image was then used as a compensation filter. The results show that SI is reduced, BER characteristics are improved, and communication quality is improved.